US6018752A - Microprocessor for performing unsigned operations with signed instructions - Google Patents
Microprocessor for performing unsigned operations with signed instructions Download PDFInfo
- Publication number
- US6018752A US6018752A US08/876,699 US87669997A US6018752A US 6018752 A US6018752 A US 6018752A US 87669997 A US87669997 A US 87669997A US 6018752 A US6018752 A US 6018752A
- Authority
- US
- United States
- Prior art keywords
- unsigned
- data
- signed
- microprocessor
- instruction
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F7/00—Methods or arrangements for processing data by operating upon the order or content of the data handled
- G06F7/38—Methods or arrangements for performing computations using exclusively denominational number representation, e.g. using binary, ternary, decimal representation
- G06F7/48—Methods or arrangements for performing computations using exclusively denominational number representation, e.g. using binary, ternary, decimal representation using non-contact-making devices, e.g. tube, solid state device; using unspecified devices
- G06F7/50—Adding; Subtracting
Definitions
- the present invention relates to operations of a microprocessor. More particularly, the present disclosure describes a microprocessor operable to perform signed arithmetic instructions for operations with unsigned data.
- Binary data can be formed to represent signed data or unsigned data. Signed data can be used to represent both positive and negative numbers. Unsigned data can only represent positive numbers. Arithmetic instructions of a microprocessor process information in form of signed and unsigned digital data and control many operational processes of the microprocessor.
- a system may use "signed" arithmetic instructions for some arithmetic operations which are operable only with signed numbers. Conversely, instructions for some arithmetic operations may be operable only for unsigned numbers and would thereby be referred to as "unsigned" arithmetic instructions.
- vis 13 pdist in the Visual Instruction Set (VIS) of UltraSPARC processors controls a microprocessor to compute the sum of the absolute difference between a pair of eight-pixel unsigned binary bit sets. This is described in the "VIS User's Manual", release 1.0, from Sun Microsystems, April, 1996.
- VIS User's Manual release 1.0
- vis 13 pdist instruction usually can be used for unsigned data only. Its current implementation cannot be directly applied to signed data.
- Other examples of instructions which can only be applied to signed data are also known, such as the instructions fpsub16 and pfsub32 on UltraSPARC processors.
- the inventor recognized that many applications require operations for both signed and unsigned data. In view of this recognition, the inventor has discovered techniques allowing implementing a signed arithmetic instruction on unsigned data.
- the inventor has discovered additional techniques allowing implementing an unsigned instruction on signed data.
- microprocessor the capabilities of a microprocessor are enhanced by controlling the microprocessor in a special way so that a signed arithmetic instruction can be used on unsigned data.
- the microprocessor is also controlled such that an unsigned instruction can be used on signed data.
- One aspect of the present invention is an implementation of a data transforming mechanism in a microprocessor which renders the instructions thereof transparent to the type of data.
- Another more specific aspect of the present invention is to expand the functionality of the Visual Instruction Set (VIS), thereby enhancing the performance of UltraSPARC processors.
- VIS Visual Instruction Set
- a preferred microprocessor for performing signed and unsigned operations implements a data transformation mechanism that comprises:
- FIG. 1 is a block diagram showing an exemplary microprocessor.
- FIG. 2 is a flowchart of a preferred operation of transforming an unsigned operation between two unsigned integers to an equivalent operation between two signed integers using signed instructions.
- FIG. 3 is a flowchart of a preferred operation of transforming a signed operation between two signed integers to an equivalent operation between two unsigned integers using unsigned instructions.
- the present invention implements a mechanism in a microprocessor to transform an item of data between signed and unsigned formats so that the data can be processed by both signed and unsigned instructions.
- a mechanism of performing transformation between signed and unsigned formats for operations on subtraction is disclosed in the present disclosure.
- an operation between two signed integers can be transformed into another equivalent operation between two unsigned integers as needed.
- an operation between two unsigned integers can be transformed into an equivalent operation of two signed integers. It is noted that the transformations do not change the end results of the operations.
- FIG. 1 shows a preferred microprocessor 101 in accordance with the invention.
- An arithmetic logic unit (ALU) 102 performs arithmetic and logic operations including binary addition, subtraction, multiplication, division, and a number of logical comparisons.
- the instruction set for the microprocessor 101 controls, at least in part, the operations of ALU 102.
- the instruction set can be stored in a memory unit 110 which can be either an built-in memory in the microprocessor and/or an external memory.
- the data from the data register 106 may be transformed into a desired format (e.g., from signed to unsigned) prior to an execution of an operation by the ALU 102.
- Unsigned integer data of N bits can be represented in the following binary form:
- N bits where s is the value of the most significant bit of the data and frac is the value of the lower (N-1) bits, respectively.
- a signed integer data of N bits can be analogously expressed in another form as
- Subtraction c between two unsigned data a and b can be represented as follows:
- Equation (5) can be rewritten in the following form by using two signed integers m and n having values respectively corresponding to unsigned data a and b, where: ##EQU1## Exchanging the most significant bits of a and b and adding appropriate sign bit to each as shown in Equations (6) and (7) produce two new signed integers m and n. Note that the subtraction is carried out the same way with the same end results, but with different operands.
- Equation (8) states that an unsigned operation represented by Equation (5) can be transformed into a form that can be operated using signed instructions. The transformation includes using two signed integers m and n as new operands in a second data representation to replace the original operands in a first data representation, unsigned a and b.
- FIG. 2 is a flowchart 200 to show the main steps of transforming an unsigned operation between two unsigned integers to an equivalent operation between two signed integers using signed instructions.
- step one unsigned operands are transformed to unsigned operands.
- step two the new unsigned integers are converted into two new signed integers by adding the bit information indicative of sign.
- step two signed instructions are used to perform an operation between the two new signed integers.
- Analogous transformation can be used to transform signed operations for unsigned instructions.
- Subtraction z between two signed data x and y with the same bit pattern can be expressed as: ##EQU2## Exchanging the most significant bits of x and y in the following manner defines two unsigned integers p and q that can be used to replace x and y in Equation (11): ##EQU3## This exchange of the most significant bits allows a transformation from a first data representation, signed x and y, to a second data representation, unsigned p and q, as defined by Equations (12) and (13). Such a conversion hence enables the microprocessor to perform unsigned instructions on signed operations.
- FIG. 3 is a flowchart 300 to show the main steps of transformation of an operation between two signed integers to an equivalent operation between two unsigned integers, using unsigned instructions.
- step one signed operands are transformed to unsigned operands.
- step two the new signed integers are converted into two new unsigned integers by eliminating the bit information indicative of sign.
- step two unsigned instructions are used to perform an operation between the two new unsigned integers.
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Pure & Applied Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computing Systems (AREA)
- Mathematical Optimization (AREA)
- General Engineering & Computer Science (AREA)
- Executing Machine-Instructions (AREA)
Abstract
A microprocessor operable to transform unsigned data to a signed format so that the unsigned data can be processed by signed instructions. In particular, a subtraction between two unsigned numbers can be transformed into a subtraction between two signed numbers.
Description
The present invention relates to operations of a microprocessor. More particularly, the present disclosure describes a microprocessor operable to perform signed arithmetic instructions for operations with unsigned data.
Operations of digital microprocessors are based on binary data. Binary data can be formed to represent signed data or unsigned data. Signed data can be used to represent both positive and negative numbers. Unsigned data can only represent positive numbers. Arithmetic instructions of a microprocessor process information in form of signed and unsigned digital data and control many operational processes of the microprocessor.
Many prior-art systems have separate operations for different kinds of information. For example, a system may use "signed" arithmetic instructions for some arithmetic operations which are operable only with signed numbers. Conversely, instructions for some arithmetic operations may be operable only for unsigned numbers and would thereby be referred to as "unsigned" arithmetic instructions.
For example, an instruction called "vis13 pdist" in the Visual Instruction Set (VIS) of UltraSPARC processors controls a microprocessor to compute the sum of the absolute difference between a pair of eight-pixel unsigned binary bit sets. This is described in the "VIS User's Manual", release 1.0, from Sun Microsystems, April, 1996. One application of this instruction is to accelerate motion compensation to support real-time video compression. However, vis13 pdist instruction usually can be used for unsigned data only. Its current implementation cannot be directly applied to signed data. Other examples of instructions which can only be applied to signed data are also known, such as the instructions fpsub16 and pfsub32 on UltraSPARC processors.
Such division between signed and unsigned operations is limiting in the operation efficiency of a microprocessor.
The inventor recognized that many applications require operations for both signed and unsigned data. In view of this recognition, the inventor has discovered techniques allowing implementing a signed arithmetic instruction on unsigned data.
The inventor has discovered additional techniques allowing implementing an unsigned instruction on signed data.
Therefore, the capabilities of a microprocessor are enhanced by controlling the microprocessor in a special way so that a signed arithmetic instruction can be used on unsigned data. The microprocessor is also controlled such that an unsigned instruction can be used on signed data.
One aspect of the present invention is an implementation of a data transforming mechanism in a microprocessor which renders the instructions thereof transparent to the type of data.
Another more specific aspect of the present invention is to expand the functionality of the Visual Instruction Set (VIS), thereby enhancing the performance of UltraSPARC processors.
A preferred microprocessor for performing signed and unsigned operations implements a data transformation mechanism that comprises:
using a data transformation to change a data representation of a first form for data that is originally in a second form, one of said first form and said second form representing a signed format and another representing an unsigned format;
transforming a second form operation of said data in said second form controlled by a second form instruction into a first form operation in said data representation of said first form; and
using said microprocessor to perform an unsigned operation on an unsigned data with a signed instruction and a signed operation on a signed data with an unsigned instruction.
These and other aspects and advantages of the present invention will become more apparent in light of the following detailed description of preferred embodiments thereof, as illustrated in the accompanying drawings, in which:
FIG. 1 is a block diagram showing an exemplary microprocessor.
FIG. 2 is a flowchart of a preferred operation of transforming an unsigned operation between two unsigned integers to an equivalent operation between two signed integers using signed instructions.
FIG. 3 is a flowchart of a preferred operation of transforming a signed operation between two signed integers to an equivalent operation between two unsigned integers using unsigned instructions.
The present invention implements a mechanism in a microprocessor to transform an item of data between signed and unsigned formats so that the data can be processed by both signed and unsigned instructions. In particular, a mechanism of performing transformation between signed and unsigned formats for operations on subtraction is disclosed in the present disclosure. According to the invention, an operation between two signed integers can be transformed into another equivalent operation between two unsigned integers as needed. Conversely, an operation between two unsigned integers can be transformed into an equivalent operation of two signed integers. It is noted that the transformations do not change the end results of the operations.
FIG. 1 shows a preferred microprocessor 101 in accordance with the invention. An arithmetic logic unit (ALU) 102 performs arithmetic and logic operations including binary addition, subtraction, multiplication, division, and a number of logical comparisons. The instruction set for the microprocessor 101 controls, at least in part, the operations of ALU 102. The instruction set can be stored in a memory unit 110 which can be either an built-in memory in the microprocessor and/or an external memory. The data from the data register 106 may be transformed into a desired format (e.g., from signed to unsigned) prior to an execution of an operation by the ALU 102.
Unsigned integer data of N bits can be represented in the following binary form:
s.2.sup.N-1 +frac, (N=1, 2, 3, . . . ) (1)
where s is the value of the most significant bit of the data and frac is the value of the lower (N-1) bits, respectively. A signed integer data of N bits can be analogously expressed in another form as
-s.2.sup.N- +frac, (2)
where the negative sign "-" in front of s indicates that the integer is a signed aperand.
1. Transforming Unsigned Operations for Signed Instructions
Subtraction c between two unsigned data a and b can be represented as follows:
a=s.sub.1.2.sup.N-1 +frac1, (3)
b=s.sub.2.2.sup.N-1 +frac2, (4)
then,
c=a-b=(s.sub.1 -s.sub.2).2.sup.N-1 +(frac1-frac2) (5)
Equation (5) can be rewritten in the following form by using two signed integers m and n having values respectively corresponding to unsigned data a and b, where: ##EQU1## Exchanging the most significant bits of a and b and adding appropriate sign bit to each as shown in Equations (6) and (7) produce two new signed integers m and n. Note that the subtraction is carried out the same way with the same end results, but with different operands. Equation (8) states that an unsigned operation represented by Equation (5) can be transformed into a form that can be operated using signed instructions. The transformation includes using two signed integers m and n as new operands in a second data representation to replace the original operands in a first data representation, unsigned a and b.
FIG. 2 is a flowchart 200 to show the main steps of transforming an unsigned operation between two unsigned integers to an equivalent operation between two signed integers using signed instructions. The above example for subtraction is only a special case. In step one, unsigned operands are transformed to unsigned operands. For subtraction, the most significant bits of the two unsigned integers are exchanged and hence two new unsigned integers are produced. Next, the new unsigned integers are converted into two new signed integers by adding the bit information indicative of sign. In step two, signed instructions are used to perform an operation between the two new signed integers.
2. Transforming Signed Operations for Unsigned Instructions
Analogous transformation can be used to transform signed operations for unsigned instructions.
Subtraction z between two signed data x and y with the same bit pattern can be expressed as: ##EQU2## Exchanging the most significant bits of x and y in the following manner defines two unsigned integers p and q that can be used to replace x and y in Equation (11): ##EQU3## This exchange of the most significant bits allows a transformation from a first data representation, signed x and y, to a second data representation, unsigned p and q, as defined by Equations (12) and (13). Such a conversion hence enables the microprocessor to perform unsigned instructions on signed operations.
FIG. 3 is a flowchart 300 to show the main steps of transformation of an operation between two signed integers to an equivalent operation between two unsigned integers, using unsigned instructions. The above example for subtraction is only a special case. In step one, signed operands are transformed to unsigned operands. For subtraction, the most significant bits of the two signed integers are exchanged and hence two new signed integers are produced. Next, the new signed integers are converted into two new unsigned integers by eliminating the bit information indicative of sign. In step two, unsigned instructions are used to perform an operation between the two new unsigned integers.
Although the present invention has been described in detail with reference to the preferred embodiment, one ordinarily skilled in the art to which this invention pertains will appreciate that various modifications and enhancements may be made without departing from the spirit and scope of the following claims.
Claims (3)
1. A microprocessor capable of
forming a signed representation of unsigned data and performing an operation on two unsigned data by using an instruction for signed data, comprising:
means for separating the most significant bit from other bits of each of first and second unsigned data;
means for combining the most significant bit of said first unsigned data to the other bits of said second unsigned data and adding a sign bit to produce a first signed data;
means for combining the most significant bit of said second unsigned data to the other bits of said first unsigned data and adding a sign bit to produce a second signed data,
wherein a selected operation on first and second unsigned data requires an unsigned instruction designed for operating on unsigned data; and
an operating means for applying a selected signed instruction on said first and second signed data so as to effect said unsigned instruction on said first and second unsigned data without requiring said unsigned instruction in the microprocessor.
2. A microprocessor as in claim 1,
wherein said selected operation includes a subtraction between two unsigned numbers having the same number of significant bits; and
said substraction between said two unsigned numbers is performed by a substraction between two signed numbers generated from said two unsigned numbers.
3. A microprocessor capable of performing substraction, comprising:
means for exchanging the most significant bits of first and second unsigned data to produce third and fourth unsigned data;
means for adding a sign bit in each of said third and fourth unsigned data to respectively form a first signed data from said third unsigned data and a second signed data from said fourth unsigned data; and
an operating means for performing a signed subtraction between said first signed data and said second signed data to effect an unsigned subtraction operation between said first and second unsigned data without an instruction for said unsigned substraction.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US08/876,699 US6018752A (en) | 1997-06-16 | 1997-06-16 | Microprocessor for performing unsigned operations with signed instructions |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US08/876,699 US6018752A (en) | 1997-06-16 | 1997-06-16 | Microprocessor for performing unsigned operations with signed instructions |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| US6018752A true US6018752A (en) | 2000-01-25 |
Family
ID=25368386
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| US08/876,699 Expired - Fee Related US6018752A (en) | 1997-06-16 | 1997-06-16 | Microprocessor for performing unsigned operations with signed instructions |
Country Status (1)
| Country | Link |
|---|---|
| US (1) | US6018752A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4935890A (en) * | 1988-01-18 | 1990-06-19 | Kabushiki Kaisha Toshiba | Format converting circuit for numeric data |
| EP0489552A2 (en) * | 1990-12-03 | 1992-06-10 | I.G.E. Medical Systems Limited | Image processing system |
| US5745125A (en) * | 1996-07-02 | 1998-04-28 | Sun Microsystems, Inc. | Floating point processor for a three-dimensional graphics accelerator which includes floating point, lighting and set-up cores for improved performance |
| US5801975A (en) * | 1996-12-02 | 1998-09-01 | Compaq Computer Corporation And Advanced Micro Devices, Inc. | Computer modified to perform inverse discrete cosine transform operations on a one-dimensional matrix of numbers within a minimal number of instruction cycles |
-
1997
- 1997-06-16 US US08/876,699 patent/US6018752A/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4935890A (en) * | 1988-01-18 | 1990-06-19 | Kabushiki Kaisha Toshiba | Format converting circuit for numeric data |
| EP0489552A2 (en) * | 1990-12-03 | 1992-06-10 | I.G.E. Medical Systems Limited | Image processing system |
| US5745125A (en) * | 1996-07-02 | 1998-04-28 | Sun Microsystems, Inc. | Floating point processor for a three-dimensional graphics accelerator which includes floating point, lighting and set-up cores for improved performance |
| US5801975A (en) * | 1996-12-02 | 1998-09-01 | Compaq Computer Corporation And Advanced Micro Devices, Inc. | Computer modified to perform inverse discrete cosine transform operations on a one-dimensional matrix of numbers within a minimal number of instruction cycles |
Non-Patent Citations (2)
| Title |
|---|
| Japanese Abstract No. 362224830; Ushinmei, Oct. 1987. * |
| Japanese Abstract No. 405303498; Nakamura et al., Nov. 1993. * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR100348951B1 (en) | Memory store from a register pair conditional | |
| US5500812A (en) | Multiplication circuit having rounding function | |
| TW448400B (en) | Processor which can favorably execute a rounding process composed of positive conversion saturation calculation processing | |
| US5859997A (en) | Method for performing multiply-substrate operations on packed data | |
| US6014684A (en) | Method and apparatus for performing N bit by 2*N-1 bit signed multiplication | |
| US7395298B2 (en) | Method and apparatus for performing multiply-add operations on packed data | |
| US7430578B2 (en) | Method and apparatus for performing multiply-add operations on packed byte data | |
| US5983256A (en) | Apparatus for performing multiply-add operations on packed data | |
| JP4064989B2 (en) | Device for performing multiplication and addition of packed data | |
| EP0657803B1 (en) | Three input arithmetic logic unit | |
| US5991865A (en) | MPEG motion compensation using operand routing and performing add and divide in a single instruction | |
| RU2275677C2 (en) | Method, device and command for performing sign multiplication operation | |
| US5880979A (en) | System for providing the absolute difference of unsigned values | |
| EP0657802A2 (en) | Rotation register for orthogonal data transformation | |
| US5742529A (en) | Method and an apparatus for providing the absolute difference of unsigned values | |
| US7047269B2 (en) | Cordic method and architecture applied in vector rotation | |
| US6029184A (en) | Method of performing unsigned operations with signed instructions in a microprocessor | |
| US6018751A (en) | Microprocessor for performing signed operations with unsigned instructions | |
| US6085207A (en) | Method of performing signed operations with unsigned instructions in a microprocessor | |
| US6018752A (en) | Microprocessor for performing unsigned operations with signed instructions | |
| US5661674A (en) | Divide to integer | |
| US5907500A (en) | Motion compensation adder for decoding/decompressing compressed moving pictures | |
| US6504495B1 (en) | Clipping data values in a data processing system | |
| JP2753922B2 (en) | Fixed-point division method | |
| US5831882A (en) | Orthogonal transformation processing device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| AS | Assignment |
Owner name: SUN MICROSYSTEMS, INC., CALIFORNIA Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:HE, LIANG;REEL/FRAME:008873/0170 Effective date: 19971024 |
|
| FEPP | Fee payment procedure |
Free format text: PAYOR NUMBER ASSIGNED (ORIGINAL EVENT CODE: ASPN); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY |
|
| FPAY | Fee payment |
Year of fee payment: 4 |
|
| REMI | Maintenance fee reminder mailed | ||
| LAPS | Lapse for failure to pay maintenance fees | ||
| STCH | Information on status: patent discontinuation |
Free format text: PATENT EXPIRED DUE TO NONPAYMENT OF MAINTENANCE FEES UNDER 37 CFR 1.362 |
|
| FP | Lapsed due to failure to pay maintenance fee |
Effective date: 20080125 |